Technical Program

Paper Detail

Paper:	SLP-L4.2
Session:	Speech Synthesis I
Time:	Wednesday, May 17, 10:20 - 10:40
Presentation:	Lecture
Topic:	Speech and Spoken Language Processing: Signal Processing/Statistical Model for synthesis
Title:	HSMM-BASED MODEL ADAPTATION ALGORITHMS FOR AVERAGE-VOICE-BASED SPEECH SYNTHESIS
Authors:	Junichi Yamagishi, Katsumi Ogata, Yuji Nakano, Juri Isogai, Takao Kobayashi, Tokyo Institute of Technology, Japan
Abstract:	In HMM-based speech synthesis, we have to choose the modeling strategy for speech synthesis unit depending on the amount of available speech data to generate synthetic speech of better quality. In general, speaker-dependent modeling is an ideal choice for a large speech data, whereas speaker adaptation with average voice model becomes promising when available speech data of a target speaker is limited. This paper describes several speaker adaptation algorithms and MAP modification to develop consistent method for synthesizing speech in a unified way for arbitrary amount of the speech data. We incorporate these adaptation algorithms into our HSMM-based speech synthesis system and show its effectiveness from results of several evaluation tests.